Memory for specific visual details can be enhanced by negative ...

More documents

Recommendations

Info

108 E.A. Kensinger et al. / Journal of Memory and Language 54 (2006) 99–112 tions between response type and presentation rate (F(2,46) = 3.78, p < .05, partial eta-squared = .14), response type and emotion type (F(2,46) = 8.71, p < .001, partial eta-squared = .28), and among response type, presentation rate, and emotion type (F(2,46) = 3.49, p < .05, partial eta-squared = .13). We conducted separate ANOVAs for the items studied at 500 and 1000 ms. For the items with a 500 ms presentation rate, the ANOVA indicated a significant main effect of response type (F(2,46) = 5.53, p < .01, partial eta-squared = .19) and an interaction between response type and emotion (F(2,46) = 4.13, p < .02, partial etasquared = .15). For the items with a 1000 ms presentation rate, the ANOVA also indicated a significant main effect of response type (F(2,46) = 20.57, p < .001, partial eta-squared = .47) and an interaction between response type and emotion type (F(2,46) = 9.77, p < .001, partial eta-squared = .30). Post hoc t tests confirmed that for items studied at both the 500 and 1000 ms presentation rates, individuals were more likely to later correctly recognize the same items as ‘‘same’’ if they were emotional than if they were neutral (t(23) = 2.82, p < .01 for 500 ms presentation; t(23) = 3.71, p < .001 for 1000 ms presentation). For the 500 ms presentation rate, the results replicated those of Experiment 1, in that the interaction between response type and emotion type was driven by the change in the specific recognition responses, and not in the general recognition of the objects: Sums of ‘‘same’’ and ‘‘similar’’ responses to same items did not differ based on the emotion type (t(23) < 1.4). For items studied for 1000 ms, however, not only did emotion enhance specific recognition, it also increased general recognition. The sums of ‘‘same’’ and ‘‘similar’’ responses were greater for same emotional items than for same neutral items (t(23) = 3.09, p < .01; Table 2). The results discussed so far for the same items suggested that for items presented for either 500 or 1000 ms, emotional content confers a memory advantage in terms of increasing the specific recognition responses. To further examine the validity of this conclusion, we calculated the probability that participants showed specific recognition of a same item, given that they had remembered at least some aspects of the item (i.e., the probability that a same item was labeled ‘‘same,’’ given that it was labeled ‘‘same’’ or ‘‘similar’’ [not ‘‘new’’]). We then conducted an ANOVA on these proportions, with presentation rate and emotion as within-subject factors. This ANOVA revealed main effects of presentation rate (F(1,23) = 5.65, p < .05, partial eta-squared = .20) and emotion (F(1,23) = 7.77, p < .01, partial eta-squared = .25) as well as an interaction between the two (F(1,23) = 7.17, p < .05, partial eta-squared = .24). This interaction reflected the fact that, when individuals showed any recognition of an item (assigning a ‘‘same’’ or ‘‘similar’’ response to a same item), a greater proportion of those recognition responses were based on specific recognition for the emotional items than for the neutral items (t(23) = 2.45, p < .05 for items studied at 500 ms; t(23) = 2.89, p < .01 for items studied at 1000 ms 3 ). There were no interactions with presentation rate. Thus, given that at least some aspects of an item were remembered, presentation rate did not affect the likelihood that specific visual information was available. Following both 500 and 1000 ms presentation durations, emotional items that were remembered were more likely to be recognized with specific visual details than were neutral items that were remembered (Fig. 2). Discussion The central question asked in Experiment 2 was whether the effects of emotion that existed when objects were studied for 500 ms (i.e., increased specific recognition) would remain unchanged when items were studied for a longer (1000 ms) duration, or whether the effects of emotion would be altered when items were studied for a longer period of time. The results of Experiment 2 indicate that emotional content continues to enhance specific recognition, even when items are studied for 1000 ms. Moreover, with the longer presentation rate, emotional content also seems to affect general recognition: Individuals were more likely to remember at least some aspects of items (i.e., that they had seen a particular type of item) if it was an emotional object presented for 1000 ms than if it was a neutral object presented for 1000 ms. Prior research has been mixed with regard to whether or not emotion enhances the likelihood that any information is recognized about an item (e.g., overall recognition scores; Abrisqueta-Gomez, Bueno, Oliveira, & Bertolucci, 2002; Johansson, Mecklinger, & Treese, 2004; Kensinger & Corkin, 2003; Kensinger & Schacter, in press; Windmann & Kutas, 2001) or only the likelihood that information is remembered vividly (e.g., Ochsner, 2000; Sharot et al., 2004). The present results indicate that both types of memory benefits can occur (see also Burke et al., 1992): When items are processed for a longer period of time, emotional content appears to enhance not only the amount of specific information (or visual detail) remembered about an item, but also the likelihood that any information (specific or general) will be remembered. 3 As in Experiment 1, these effects did not reflect a general bias for participants to assign ‘‘same’’ responses to emotional objects. Emotion did not affect the distribution of ‘‘same’’ responses given to similar items (t(23) < 1 for items studied at 500 and 1000 ms).
E.A. Kensinger et al. / Journal of Memory and Language 54 (2006) 99–112 109 General discussion Abundant evidence indicates that individuals believe that they remember emotional information vividly (e.g., Dewhurst & Parry, 2000; Kensinger & Corkin, 2003; Ochsner, 2000; Reisberg et al., 1988). It has been debated to what extent these self-report data reflect accurately on the detail with which emotional information is remembered (e.g., Easterbrook, 1959; Loftus & Burns, 1982; Sharot et al., 2004). While previous studies have demonstrated that negative arousing content can enhance the likelihood that various aspects of an event are remembered (e.g., the spatial location of a word, whether an object was visually presented or mentally imagined; DÕArgembeau & Van der Linden, 2004; Doerksen & Shimamura, 2001; Kensinger & Schacter, in press; MacKay & Ahmetzanov, 2005), the results have been ambiguous with regard to whether emotion enhances memory for an itemÕs specific visual details, with some evidence indicating an enhancement (e.g., Burke et al., 1992) and other data suggesting a decrement (e.g., Adolphs et al., 2001). The present results provide strong evidence that negative arousing content can increase the likelihood that the specific visual details of an emotional object are remembered. A few factors may contribute to this enhancement. For one, emotional arousal has strong effects on attention deployment: Attention appears to be focused automatically and preferentially on emotionally arousing stimuli, increasing the likelihood that emotional items are perceived. For example, individuals are more likely to detect emotional objects in complex arrays than they are to detect nonemotional ones (Ohman, Flykt, & Esteves, 2001), and in cuing paradigms, emotional cues are more effective than neutral ones in directing participantsÕ attention to a particular location on a screen (Armony & Dolan, 2002; Mogg, Bradley, de Bono, & Painter, 1997). Moreover, emotionally arousing items appear privy to prioritized or facilitated processing, such that emotional items can be processed even when attention is limited or divided (reviewed by Dolan & Vuilleumier, 2003). Thus, emotional items are less susceptible to the attentional blink than are nonemotional ones (Anderson & Phelps, 2001), and patients with visual extinction are less likely to ignore an emotional stimulus presented to the contralesional visual field than they are to ignore a nonemotional stimulus (e.g., Vuilleumier & Schwartz, 2001). These attentional effects of emotion may relate to the enhanced memory for visual details demonstrated here. The encoding of visual details may have occurred more automatically for the emotional items than for the neutral items. This hypothesis is consistent with neuroimaging data indicating that visual activity is greater during the processing of emotional information compared to neutral information (Lang et al., 1998; Vuilleumier, Richardson, Armony, Driver, & Dolan, 2004; Wang, McCarthy, Song, & LaBar, 2005) and that these effects can occur even with limited attention (Morris, Ohman, & Dolan, 1999; Vuilleumier, Armony, Driver, & Dolan, 2001; Lane, Chua, & Dolan, 1999). This attentional focus could have enhanced memory for visual details in another way as well: For emotional items, a greater proportion of time may have been spent attending to the visual stimulus rather than to other types of information (either nonvisual details related to the presented item, or task-irrelevant information). As discussed in the introduction, emotional elements in a scene often appear to be attended at the cost of information present in the periphery (e.g., Brown, 2003; Burke et al., 1992; Kensinger et al., 2005; Safer et al., 1998). This focusing will mean that encoding resources will be more likely to be focused on the visual percept of the emotional item than on other information in the environment. Although the stimuli used in the present experiment were single objects, and therefore attention did not have to be diverted from other visual elements presented on the computer screen, it is still possible that other aspects of the testing environment (including thoughts associated with the types of items presented) were more likely to distract attention from encoding of the visual form of the neutral items compared to the emotional items. It is even possible that these differences in attention deployment could explain why the effects of emotion were most robust with longer presentation durations. At 250 ms, participantsÕ attention may have been focused on the visual form of the neutral or emotional items for their entire presentation duration. With longer durations, in contrast, it is possible that attention began to divert from the visual percept, but that this diversion occurred more frequently for the neutral items than for the emotional items. Such a diversion of attention may also explain why the longer presentation rates boosted specific recognition significantly for the emotional stimuli but had less of an effect on specific recognition for the neutral stimuli: Participants may not have attended to the visual percept of the neutral items for the entire time that they were on the computer screen, whereas such focal attention may have been more successful for the emotional items. Future studies will be required to investigate the generality of this finding; it is currently unclear whether sufficient study time is required principally when encoding of specific visual details is required, or whether insufficient study duration can reduce emotional memory enhancement effects across a range of paradigms. This attentional explanation may describe how the emotional content of the items affected specific recognition through effects at encoding. It is well known, however, that the emotional content of information also can influence other stages of memory processing (consolidation and retrieval). For example, memory enhancements
Page 1 and 2: Journal of Memory and Language 54 (
Page 3 and 4: E.A. Kensinger et al. / Journal of
Page 9: E.A. Kensinger et al. / Journal of

Memory for specific visual details can be enhanced by negative ...

Create successful ePaper yourself

Delete template?

Save as template?